The overall hypothesis of this application is that white matter (WM), especially the limbic tracts, is one of the primary targets of Alzheimer's disease (AD) and diffusion tensor imaging (DTI) can sensitively detect changes in these WM tracts. The goal is to develop quantitative DTI image analysis techniques to detect WM abnormalities in AD. AD is the most common cause of dementia. Although the primary pathology is cortical neuronal cell degeneration, increasing evidence indicates a preponderance of WM pathology over gray matter pathology. Moreover, WM alterations could be an indirect indicator of nerve cell loss, since the volume of a nerve cell is much smaller than its myelinated fiber. Therefore, WM seems to be a good focus for both early diagnosis and monitoring of disease progression. MRI is a non-invasive technique that is widely available in the United States, which has great potential as an imaging biomarker. However, conventional MRI cannot provide contrasts to differentiate various WM structures and has low sensitivity and specificity for detecting changes in specific WM structures. DTI is a method that has the potential to detect abnormalities in specific white matter structures. Using this method can increase the sensitivity and specificity to detect WM abnormalities, compared to conventional MRI analysis. However, quantification techniques for DTI data have not been well-developed and it has been difficult to fully exploit the WM anatomical information revealed by DTI. This application is based on two technical innovations we have developed: one is a white matter brain atlas (JHU-DTI-MNI) in stereotaxic coordinates that contains detailed white matter maps based on diffusion tensor imaging;and the other is the state-of-the-art computational neuroanatomy technology based on the highly-elastic non-linear brain normalization method (LDDMM), which can preserve WM fiber connectivity in the transformation process. In Aim 1, we will extend this effort to 1) optimize the white matter atlas for the elderly population, and 2) test our advanced cost functions of LDDMM to improve the normalization quality. After the optimization, we will apply these techniques to detect WM abnormalities in AD and Mild Cognitive Impairment (MCI) patients by cross-sectional analysis (Aim 2) and longitudinal analysis (Aim 3). An existing longitudinal MRI and clinical database acquired at Johns Hopkins University is available for this study. The data were acquired every four months for one year from the same subjects. In Aim 2, normalized DTI data from AD and MCI patients were compared to age-matched controls to detect disease-specific alterations in morphology and MR parameters (diffusion constant, diffusion anisotropy, and T2). In Aim 3, we will characterize disease progression-related changes in morphology and MR parameters by a longitudinal analysis. In summary, we propose a new image-analysis method for a comprehensive WM survey that will detect disease-specific changes and disease progression-specific changes in MCI and AD. PUBLIC HEALTH RELEVANCE: We will develop new imaging biomarkers for Alzheimer's disease. Our image-analysis method based on probabilistic white matter atlas and white matter fiber direction-oriented transformation enable us to comprehensively survey the white matter structures to detect disease specific and time-dependent alteration of the brain.